Advanced Feature Store Concepts MACHINE-LEARNING-PROFESSIONAL Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   What is the purpose of configuring an online table in Databricks?

   • ATo provide a backup of feature tables for disaster recovery.
   • BTo replace the need for a feature store by storing all features in a single table.
   • CTo store historical feature values for batch training jobs.
   • DTo provide low-latency access to feature values for real-time model serving.
   Reveal Solution Discussion Select an answer first
2. 2foundation · easy

   What is the primary benefit of using streaming for feature ingestion in a feature store?

   • AIt allows features to be updated in near real-time, enabling models to use the most current data for inference.
   • BIt reduces the storage cost of the feature store by compressing data.
   • CIt guarantees exactly-once processing of all data.
   • DIt eliminates the need for batch processing entirely.
   Reveal Solution Discussion Select an answer first
4. 3foundation · easy

   What is the role of the FeatureEngineering Client in Databricks?

   • ATo train machine learning models using feature store data.
   • BTo serve features to online models via REST endpoints.
   • CTo monitor model performance in production.
   • DTo create and manage feature tables and orchestrate automated feature computation pipelines.
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A gaming company wants to serve real-time features for a recommendation model. They have a Delta table with player features that is updated every 10 minutes via a streaming pipeline. They need to configure an online table for low-latency serving. Which Databricks SDK call should they use?

   • Acreate_pipeline with the Delta table path.
   • Bcreate_serving_endpoint with the Delta table path.
   • Ccreate_online_table with the source Delta table path and primary key.
   • Dcreate_feature_table with the Delta table path.
   Reveal Solution Discussion Select an answer first
6. 5foundation · easy

   How does Databricks Feature Serving ensure consistency between training and production for on-demand features?

   • ABy storing the on-demand feature values in the feature store for both training and inference.
   • BBy ensuring that the model is retrained every time the on-demand feature logic changes.
   • CBy allowing you to define the on-demand feature computation once and reuse the same function in both training and serving.
   • DBy using the same online table for both training and inference.
   Reveal Solution Discussion Select an answer first